This note discusses the processing that was applied at NSSDC in June 2005 to
the Apollo 15 and Apollo 17 Heat Flow Experiment data sets PSPG-00093 and
PSPG-00022. The original data sets are:
71-063C-06A (PSPG-00093) for Apollo 15
72-096C-01A (PSPG-000222) for apollo 17

The new reformatted data sets are:
PSPG-00752
Apollo 15 Heat Flow - Thermal Conductivity Data, Reformatted to ASCII

PSPG-00753
Apollo 17 Heat Flow - Thermal Conductivity Data, Reformatted to ASCII

The documentation for the original data sets is given in Apollo15&17.Doc.pdf.

(There is also an error analysis data set: 72-096C-01B, but it is actually a 
document and is not addressed here. ) 

The data sets each contain data from two different probes associated with each
instrument. All data from a given probe are together in one file as archived
in the original data sets noted above. Each "file" as mentioned in the documentation for the original data set is actually a group of physical records that make up only a part of the original archived file. The physical records are written blocked, with 100 logical records per physical record. Where the data didn't fill a complete physical record at the end of a group, the original data had
zero values inserted to fill out the 100-logical-record physical record. The
logical record size (in 4-byte words) is either 3 words (for files 1, 2, 4,
and 5) or 7 words (for file 3). 

To make the data easier to use, NSSDC separated these groups of physical 
records into separate files (in the usual meaning of a file) by selecting the
appropriate physical records from the archived data file. Thus each original 
binary archive file resulted in five new ASCII data files, matching the
descriptions of the "files" as described in the original data set
documentation. In some cases, the number of physical records in the "file" was
found to be one less than indicated in the documentation. The new output files
are written in ASCII, unblocked, and the records containing only fill values
have been deleted. 

In the NSSDC processing program, written by H. K. Hills, a physical record of
binary data was read in, then the order of the four-byte words in the record
was inverted, making the records advance in chronological order (instead of
last to first as in the original archived data set). Then the byte order
(within a 4-byte word) was inverted to put the bytes together properly in the
VAX word (because of the way the VAX handles input tape records). Then the sign, mantissa, and characteristic of the values were selected as bit strings and moved into their own separate computer words, after which the value of each data word was computed. In cases where the input word value was zero, the output value was set to -9999 (or -9.999000E03). The logical three-word records or seven-word records were written out in ASCII directly to disk files with no blocking. Trailing filler records of zero values were deleted. The output files were named following the tape designation given in the documentation: 
  A15P1.File1 through A15P1.File5 for what the documentation calls A15P1#3
  A15P2.File1 through A15P2.File5 for A15P2#3
  A17P1.File1 through A17P1.File5 for A17P1#2
  A17P2.File1 through A17P2.File5 for A17P2#2

Thus there are now ten data files in ASCII for Apollo 15 and ten for Apollo 17, 
instead of two files for each Apollo mission. In addition to the data files, there are also a file containing the Fortran program that did the conversion of the data, and a PDF file of the relevant part of the documentation of the original data sets. This readme file is also included in the set of files that constitute the data set and that are available online on nssdcftp.gsfc.nasa.gov/spacecraft_data/apollo15_alsep/  and ... apollo17_alsep/.

These new ASCII output data files have the records in nominal chronological 
order (i.e., the inverse of the order on the original tapes), but there are 
exceptions where the time of a given point is slightly less than the time of 
the previous point. The error occurs in the 6th to the 8th significant 
digit of the time. No such time irregularities were seen in most files, but
there were numerous such cases in the files identified as A15p1.file_3 and
A17p1.fil3_3. For A15p2.file_3, there was only one. No attempt was made to
remove these time discrepancies. 


June 8, 2005
Revised April 17, 2008
Revised March 21, 2009
H. Kent Hills
